Data

Ideas

Snippets

La théorie du capital humain montre que par un investissement continu durant la carrière, les compétences s’accroissent tout au long de la carrière professionnelle et cette expérience accumulée entraîne un salaire croissant dans une perspective de cycle de vie. C’est l’expérience et non l’âge en tant que tel qui est rémunérée.

Disparities in gross hourly earnings within a country can be measured using deciles, and in particular the lowest and highest deciles, which correspond to the 10% of employees earning the least (D1) and to the 10% earning the most (D9). As a consequence, a high D9/D1 interdecile ratio indicates large disparities. D1 is the maximum gross hourly earnings received by the 10% of employees earnings least. D9 is the minimum gross hourly earnings received by the 10% of employees earning most.

Explore dataset

eu.dat$age <- factor(eu.dat$age, levels = c("Y_LT30", "Y30-39", "Y40-49", "Y50-59", "Y_GE60", "TOTAL"))
dd <- eu.dat %>% filter(indic_se == 'ERN_MN_PPS', nace_r2 == 'B-S_X_O', time == 2014) %>% 
  select(-time, -nace_r2, -indic_se)

dd1 <- eu.dat %>% filter(indic_se %in% c('ERN_MN_EUR', 'ERN_MN_PPS', 'ERN_D1_PPS', 'ERN_D9_PPS'), !geo %in% c('EA18', 'EA19'),  age == "TOTAL", nace_r2 == 'B-S_X_O', isco08 == "TOTAL", age == "TOTAL", time == 2014)
dd1 %<>% spread(indic_se, values) %>% mutate(ratio = ERN_D9_PPS / ERN_D1_PPS) %>% 
  select(-nace_r2, -isco08, -age, -time, -ERN_D9_PPS, -ERN_D1_PPS) %>%
  gather(indic_se, values, -geo)

ggplot(dd1) + geom_bar(aes(x = reorder(geo, values), y = values), stat = "identity", fill = swi_col[1]) + facet_wrap(~ indic_se, scales = "free_x") + 
  coord_flip() + theme_ipsum_rc() + scale_fill_ipsum()

## salary % diff with country average by isco-08

# add tot
dd2 <- dd %>% group_by(geo, age) %>% mutate(tot.isco = values[which(isco08 == "TOTAL")]) %>% ungroup() %>% filter(isco08 != "TOTAL")
dd2$val <- ((dd2$values - dd2$`tot.isco`) / dd2$`tot.isco`) * 100

dd2 %<>% filter(geo %in% iso2.sub)
dd2$isco08.full <- factor(isco.dic[match(dd2$isco08, isco.dic$code_name), 2] %>% unlist(use.names = F), 
                         levels = isco.dic[match(levels(dd2$isco08), isco.dic$code_name), 2] %>% unlist(use.names = F))

ggplot(dd2 %>% filter(age == "TOTAL")) + geom_tile(aes(x = geo, y = isco08.full, fill = val)) + theme_ipsum_rc() + 
  scale_fill_gradient2(high = "#193442", mid = "white", low = "#72302f", midpoint = 0)

ggplot(dd2) + geom_tile(aes(x = geo, y = isco08.full, fill = val)) + facet_wrap(~age) + theme_ipsum_rc() + 
  scale_fill_gradient2(high = "#193442", mid = "white", low = "#72302f", midpoint = 0)

## salary by age
dd3 <- dd %>% filter(geo %in% iso2.sub) %>% 
  group_by(isco08, geo) %>% arrange(age) %>% 
  mutate(wage.lt30 =  values[1])  %>% ungroup() %>% 
  filter(age != "TOTAL")
dd3$val <- (((dd3$values - dd3$wage.lt30) / dd3$wage.lt30)  * 100) 
dd3$isco08.full <- factor(isco.dic[match(dd3$isco08, isco.dic$code_name), 2] %>% 
                            unlist(use.names = F), levels = isco.dic[match(levels(dd3$isco08), isco.dic$code_name), 2] %>% unlist(use.names = F))

dd3$size <- ifelse(dd3$geo %in% c('CH', 'EU28'), 0.5, 0.1)

ggplot(dd3) + geom_line(aes(x = age, y = val, group = geo, colour = geo, size = size)) + 
  facet_wrap(~isco08.full, nrow = 3) + theme_ipsum_rc() + 
  scale_fill_gradient2(high = "#193442", mid = "white", low = "#72302f", midpoint = 0) + geom_hline(yintercept = 0) + 
  scale_colour_ipsum() + scale_size_identity() + ggtitle("% difference in average wage by age and occupation groups")

ggplot(dd3 %>% filter(isco08.full != "Managers")) + geom_line(aes(x = age, y = val, group = geo, colour = geo, size = size)) + 
  facet_wrap(~isco08.full, nrow = 3) + theme_ipsum_rc() + scale_size_identity() +
  scale_fill_gradient2(high = "#193442", mid = "white", low = "#72302f", midpoint = 0) + scale_colour_ipsum() +
  ggtitle("% difference in average wage by age and occupation groups")

## interdecile ratio and mean by median
dd4 <- eu.dat %>% filter(indic_se %in% c('ERN_D1_PPS', 'ERN_D9_PPS', "ERN_MD_PPS", "ERN_MN_PPS"), nace_r2 == 'B-S_X_O', time == 2014) %>% 
  select(-time, -nace_r2) %>% 
  spread(indic_se, values) %>% filter(geo %in% iso2.sub)

dd4$ratio <- dd4$ERN_D9_PPS / dd4$ERN_D1_PPS
dd4$ratioMeanMedian <- dd4$ERN_MN_PPS / dd4$ERN_MD_PPS

dd4$isco08.full <- factor(isco.dic[match(dd4$isco08, isco.dic$code_name), 2] %>% 
  unlist(use.names = F), levels = isco.dic[match(levels(dd4$isco08), isco.dic$code_name), 2] %>% unlist(use.names = F))

dd4$size <- ifelse(dd4$geo %in% c('CH', 'EU28'), 0.5, 0.1)

# ggplot(dd4 %>% filter( age == "TOTAL", geo != "EU28", isco08 == 'TOTAL')) + 
#   geom_bar(aes(x = reorder(geo, ratio), y = ratio), stat = "identity") + 
#   theme_ipsum_rc() + 
#   ggtitle("interdecile ratio")
# 
# ggplot(dd4 %>% filter( age == "TOTAL", geo != "EU28", isco08 == 'TOTAL')) + 
#   geom_bar(aes(x = reorder(geo, ratioMeanMedian), y = ratioMeanMedian), stat = "identity") + 
#   theme_ipsum_rc() + 
#   ggtitle("ratioMeanMedian")

dd5 <- gather(dd4, indicator, value, -isco08, -age, -geo, -ratio,-isco08.full, -size)
dd5 <- dd4
dd5 %<>% filter(geo == "CH", age != "TOTAL")
ggplot(dd5, aes(x = age, y = ERN_MN_PPS, group = isco08)) + 
  geom_line() + 
  geom_ribbon(aes(ymin = ERN_D1_PPS, ymax = ERN_D9_PPS ), fill = swi_col[2], alpha = 0.2) +
  facet_wrap(~isco08) +
  theme_ipsum_rc()

# ggplot(dd4 %>% filter( age == "TOTAL", geo != "EU28")) + geom_tile(aes(x = geo, y = isco08.full, fill = ratio)) + theme_ipsum_rc() + 
#   scale_fill_gradient2(high = "#193442", low = "#72302f") 

# ggplot(dd4 %>% filter( age == "TOTAL", geo != "EU28", isco08 == 'TOTAL')) + 
#   geom_bar(aes(x = reorder(geo, ratio), y = ratio), stat = "identity") + theme_ipsum_rc() + 
#   scale_fill_gradient2(high = "#193442", low = "#72302f")
# 
# ggplot(dd4) + geom_line(aes(x = age, y = ratio, group = geo, colour = geo, size = size)) + 
#   facet_wrap(~isco08.full, nrow = 3) + theme_ipsum_rc() + 
#   scale_colour_ipsum() + scale_size_identity() + ggtitle("% difference in average wage by age and occupation groups")

dd6 <- egais %>% filter(geo %in% iso2.sub)

dd6 %<>% group_by(geo) %>% mutate(total = values[isco08 == 'TOTAL']) %>%
  ungroup() %>% mutate(sharei = (values/total) * 100) %>% 
  filter(!isco08 %in% c("TOTAL", "NRP"))

#dd6 %>% group_by(geo) %>% mutate(test = sum(sharei)) %>% ungroup()
ggplot(data = dd6) + geom_bar(aes(x = isco08, y = sharei, fill = isco08), stat = "identity") +
  facet_wrap(~ geo) + theme_ipsum()

dd7 <- dd6 %>% filter(isco08 %in% c('OC1', 'OC2', 'OC3'))
dd7 %<>% group_by(geo) %>% mutate(or = sum(sharei)) %>% ungroup() %>%
  mutate(geo = reorder(geo, -or))

ggplot(data = dd7) +
  geom_bar(aes(x = geo, y = sharei, fill = isco08), stat = "identity") + 
  theme_ipsum()

dd7dw <- dd7 %>% select(-values, -total, -or) %>%
  spread(geo, sharei)
write.csv(dd7dw, "input/shareISCO08_OC1_OC3_dw.csv", row.names = F)

# compute mean wage for different weigths
dd8 <- dd %>% filter(age == "TOTAL", geo %in% iso2.sub) %>% select(-age)
dd8 <- dd6 %>% arrange(isco08, geo) %>% select(-values, -total) %>%
  right_join(dd8 %>% arrange(isco08, geo))
## Joining, by = c("isco08", "geo")
## Warning in right_join_impl(x, y, by$x, by$y, suffix$x, suffix$y): joining
## factors with different levels, coercing to character vector

## Warning in right_join_impl(x, y, by$x, by$y, suffix$x, suffix$y): joining
## factors with different levels, coercing to character vector
dd8 <- dd8 %>% filter(isco08 != 'TOTAL') %>% 
  group_by(isco08) %>% mutate(eu28w = sharei[geo == 'EU28'] ) %>% ungroup()
  
dd8 %>% group_by(geo) %>% 
  summarise(meanTest = weighted.mean(values, sharei, na.rm = T), 
            euWWage =  weighted.mean(values, eu28w, na.rm = T)) %>% 
  ungroup() %>%
  mutate(diff = ((euWWage - meanTest) / meanTest) * 100)
## # A tibble: 8 × 4
##     geo meanTest  euWWage       diff
##   <chr>    <dbl>    <dbl>      <dbl>
## 1    CH 4211.630 3943.140 -6.3749676
## 2    DE 3207.283 3211.868  0.1429525
## 3    ES 2121.903 2245.586  5.8288526
## 4  EU28 2551.685 2551.685  0.0000000
## 5    FR 2506.742 2464.371 -1.6902496
## 6    IT 2409.089 2552.691  5.9608619
## 7    PT 1611.980 1659.291  2.9349756
## 8    UK 2611.833 2440.885 -6.5451325
# get the non-aggregated nace_r2 categories 
naceudd8 <- dd %>% filter(age == "TOTAL", geo %in% iso2.sub) %>% select(-age)
dd8 <- dd6 %>% arrange(isco08, geo) %>% select(-values, -total) %>%
  right_join(dd8 %>% arrange(isco08, geo))
## Joining, by = c("isco08", "geo", "sharei")
## Warning in right_join_impl(x, y, by$x, by$y, suffix$x, suffix$y): joining
## factor and character vector, coercing into character vector

## Warning in right_join_impl(x, y, by$x, by$y, suffix$x, suffix$y): joining
## factor and character vector, coercing into character vector
naceu <- levels(eu.1420$nace_r2)[which(nchar(levels(eu.1420$nace_r2)) == 1)]

ee <- eu.1420 %>% filter(currency == "PPS", nace_r2 %in% c(naceu, "B-S_X_O")) %>% select(-currency)
ee$nace_r2 <- factor(ee$nace_r2, levels = c(naceu, "B-S_X_O"))
ee$nace2.full <- factor(nace.dic[match(ee$nace_r2, nace.dic$code_name), 2] %>% unlist(use.names = F), 
                         levels = nace.dic[match(levels(ee$nace_r2), nace.dic$code_name), 2] %>% unlist(use.names = F))

ee1 <- ee %>% filter(age == "TOTAL", geo %in% iso2.sub) 

ee1 %<>% group_by(geo, age) %>% mutate(tot = values[nace_r2 == "B-S_X_O"]) %>% 
  ungroup() %>% filter(nace_r2 != "B-S_X_O") %>%
   mutate(val = ((values - tot) / tot) * 100)

# order by nace_r2 by descending % change from all economic sectors for EU-28
tmp.levels <- ee1 %>% filter(geo == "EU28", age == "TOTAL") %>% 
  arrange(val) %>% select(nace2.full) %>%
  unlist(use.names = F) %>% as.character()
ee1$nace2.full <- factor(ee1$nace2.full, levels = tmp.levels)

ggplot(ee1 %>% filter(geo %in% iso2.sub, age == "TOTAL")) + geom_tile(aes(x = geo, y = nace2.full, fill = val)) + 
  theme_ipsum_rc() + 
  scale_fill_gradient2(high = "#193442", mid = "white", low = "#72302f", midpoint = 0)

ff <- eu.1423 %>% filter(currency == "PPS", nace_r2 %in% c(naceu, "B-S_X_O")) %>% select(-currency)
ff$nace_r2 <- factor(ff$nace_r2, levels = c(naceu, "B-S_X_O"))

ff$nace2.full <- factor(nace.dic[match(ff$nace_r2, nace.dic$code_name), 2] %>% unlist(use.names = F), 
                         levels = nace.dic[match(levels(ff$nace_r2), nace.dic$code_name), 2] %>% unlist(use.names = F))

ff$isced11.full <- factor(isced11.dic[match(ff$isced11, isced11.dic$code_name), 2] %>% unlist(use.names = F), 
                         levels = isced11.dic[match(levels(ff$isced11), isced11.dic$code_name), 2] %>% unlist(use.names = F))

# compute wage pc change
# ff %<>% group_by(nace_r2, geo) %>% mutate(isced.all = values[isced11 == 'TOTAL']) %>%
#   ungroup() %>%
#   mutate(val = ((values - isced.all) / isced.all) * 100)
# 
# # get overall nacer2
# ff1 <- ff %>% filter(nace_r2 == "B-S_X_O")
# 
# ggplot(ff1 %>% filter(geo %in% iso2.sub)) + geom_bar(aes(x = isced11, y = val, group = isced11, fill = isced11), stat = "identity") + 
#   theme_ipsum_rc() + facet_wrap(~ geo, ncol = 8) + 
#   theme(axis.text.x = element_text(angle = 90)) + scale_fill_ipsum()